By nikunjmavani
Single-source, multi-runtime agent tooling for core-be: skills, agents, workflow commands, and the requirement to production-ready build pipeline, all driven from agent-os/.
Reusable, project-scoped slash commands shared by all three AI tools. This
Regenerate agent-os adapters from common and fix any drift
Draft a requirement (tree first) for review, then build the production-ready slice
Run the full PR gate (pnpm ci:local) and summarize failures with fixes
Merge a PR once CI is green and approvals are in
Read-only changelog and release-notes reviewer. Scans CHANGELOG.md, recent git log, and merged PR titles to verify the changelog is accurate, complete, and follows the project's Keep a Changelog / conventional-commits format. Flags missing entries, wrong version bumps, and unreleased sections that are stale. Does NOT write commits or push — produces a gap report only.
Diagnoses a single failing core-be PR CI check and returns a short root-cause summary with a fix plan. Use when the user asks why CI failed or to diagnose a specific GitHub Actions job. Runs in isolation so verbose CI logs do not bloat the main conversation.
Runs pnpm audit, analyzes vulnerabilities, and returns a prioritized fix plan — severity, affected package, recommended action (patch/update/replace/accept). Read-only; produces a report for the user to act on, never edits package.json or lockfile.
Full on-request audit of the docs/ directory — checks index completeness, naming conventions, Mermaid diagrams, and cross-links. Use when the user asks to audit or review documentation, or after a large docs reorganization. Read-only; returns an issues list, never edits files.
Sweeps infrastructure, middleware, and config for production-hardening gaps — security headers, JWT/CORS/rate limits, DB pool/SSL, Redis, external-service resilience, logging redaction, worker limits, and CI scanning. Returns a prioritized gap list. Read-only; produces a report for the user to act on, never edits files.
Enforces the core-be public API contract conventions — snake_case route params, prefixed public ids, the uniform method→status policy, and the header naming matrix — across routes, validators, tests, OpenAPI/Postman docs, and the route-status gates. Use when adding or changing any route, param, header, public id, or response status.
Ensures code is commit-ready. Invoked when the user runs git commit (enforced by Husky pre-commit) or when the user asks to fix a failed commit. Run the guard checks and fix any failures before committing.
Maintains Toxiproxy chaos tests under src/tests/chaos/. Use when adding fault-injection scenarios or changing chaos CI/provision scripts.
Investigate a single failing PR CI check in core-be and return a short root-cause summary with a fix plan. Use when the user asks why CI failed or to diagnose a specific GitHub Actions job.
Maintains the 3-layer code quality and security pipeline (editor Biome, Husky pre-commit hooks, CI security scanning). Use after changing Biome rules, pre-commit hooks, CI workflows, lint-staged config, or adding new security tooling.
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Production-grade multi-tenant SaaS backend — Node.js · Fastify · Drizzle · BullMQ
core-be is a multi-tenant SaaS backend: a Node.js + Fastify HTTP API plus a separate BullMQ worker process, both speaking to a single Postgres database and a Redis instance. Postgres is the only source of truth — workers are pull-based, idempotent, and may be restarted without coordination.
Quick links: Quick Start · SETUP.md · Documentation · src/OVERVIEW.md · CLAUDE.md · AGENTS.md · CONTRIBUTING.md
| Layer | Technologies |
|---|---|
| Runtime / HTTP | Node.js 24, Fastify 5, Zod 4 |
| Data / Queue | Drizzle 0.45, Postgres, ioredis 5, BullMQ 5 |
| Auth / Security | jose 6, argon2id, otplib, @simplewebauthn/server, opossum |
| Integrations | Stripe 22, Resend 6, AWS SDK v3 S3 |
| Observability | Pino 10, Sentry 10, OpenTelemetry, prom-client |
| Testing / Tooling | Vitest 4, k6, Toxiproxy, Biome, pnpm 10, Husky |
Prerequisites: Node.js 24+, pnpm, Docker (for local Postgres + Redis).
pnpm install
pnpm compose:up && pnpm compose:wait
pnpm db:migrate && pnpm db:seed
pnpm dev # API on :3000
pnpm dev:worker # BullMQ worker (separate terminal)
One-command local bootstrap: pnpm setup:local (Docker + env + migrate + dev). Full clone-to-running guide: SETUP.md.
Environment: copy values from .env.example into .env.development, or run pnpm github:sync to bootstrap env files. For one-command cloud provisioning (Neon, Redis, S3, Sentry, Railway, GitHub), see setup automation (pnpm setup:infra).
One TypeScript codebase, two processes: the API (pnpm dev) and the worker (pnpm dev:worker).
flowchart LR
subgraph Clients
Browser
MobileApp[Mobile App]
Stripe
Resend
OAuth[OAuth providers]
end
subgraph API[API process: src/server.ts]
MW[middleware<br/>cors, helmet, auth, tenant,<br/>idempotency, rate-limit]
Ctrl[controllers]
Svc[services]
Repo[repositories]
EB[event-bus<br/>in-process]
end
subgraph Worker[Worker process: src/worker.ts]
Workers[BullMQ processors<br/>mail, webhook delivery,<br/>retention, reclaim]
end
subgraph Data
PG[(Postgres<br/>RLS-enforced)]
RD[(Redis<br/>BullMQ + cache)]
S3[(S3-compatible<br/>storage)]
end
Browser -->|HTTPS| MW
MobileApp -->|HTTPS| MW
Stripe -->|webhook| MW
OAuth --> MW
MW --> Ctrl --> Svc --> Repo --> PG
Svc --> EB
EB -->|enqueue| RD
Workers -->|consume| RD
Workers --> Repo
Workers --> Resend
Workers -->|customer webhooks| Browser
Repo --> PG
Svc --> S3
npx claudepluginhub nikunjmavani/core-beHarness-native ECC plugin for engineering teams - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, MCP conventions, and operator workflows for Claude Code plus adjacent agent harnesses
v9.44.1 — Patch release for Gemini environment/version detection and qwen auth gating. Run /octo:setup.
Superpowers Plus core skills library for Claude Code: planning, execution routing, TDD, debugging, and collaboration workflows
Unity Development Toolkit - Expert agents for scripting/refactoring/optimization, script templates, and Agent Skills for Unity C# development
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
Reliable automation, in-depth debugging, and performance analysis in Chrome using Chrome DevTools and Puppeteer